Selection of Grasping Target and Control System of Robotic Prosthetic Hand using Images and Deep Learning

Abstract

Robotic prosthetic hands are a device that helps to improve the quality of life for patients without hands. Recently, robotic prosthetic hands can perform various grasping patterns because of improvement of bioengineering and robotics. The research that automatically selects the appropriate operation according to the situation is important. Many previous studies have used EMG signals. However, EMG signals are difficult to generalize because EMG signals vary depending on the position of the muscle. In this study, we developed a system for controlling robotic prosthetic hands using images and deep learning to facilitate generalization. We also proposed a method for selecting a grasping target to be held in the image. These results will help to improve the quality of life of the robotic prosthetic hand user.